DOCS
← Back to Glossary
serverless
Serverless is a cloud computing model where the provider automatically provisions, scales, and bills compute on demand — so you run code, queries, and analytics without managing servers or a fixed cluster.
Serverless is a cloud computing model in which the cloud provider automatically provisions, scales, and manages the underlying servers, and you pay only for the compute you actually use. Despite the name, servers are still involved — they are simply abstracted away, so you focus on your code and queries instead of capacity planning, patching, and scaling.
How serverless works
In a serverless architecture, resources are allocated dynamically in response to demand. When a request arrives, the platform spins up compute to handle it; when traffic falls, it scales back down — often to zero, so idle workloads cost nothing. Billing is tied to actual usage (invocations, compute-seconds, or bytes scanned) rather than to provisioned, always-on capacity. The main trade-offs to know about are cold starts — a brief delay when capacity spins up from idle — and less control over the runtime environment.
The model appears at several layers of the stack:
- Functions as a service (FaaS): event-driven code such as AWS Lambda, Azure Functions, or Google Cloud Functions.
- Serverless databases and data warehouses: engines that separate storage from compute and scale query resources on demand.
- Serverless analytics: query services billed per query or per byte scanned, with no cluster to size or keep running.
Serverless databases and analytics
A serverless database decouples storage from compute so the query engine can scale independently of the data it stores, spin up on demand, and idle down when unused — removing the need to provision a fixed-size cluster. This matters most for analytical workloads, which are bursty by nature: heavy during business hours and quiet overnight. Paying only for the queries you actually run is far more efficient than keeping a warehouse running around the clock. That is what people mean by a serverless data warehouse — you load data and query it, and the platform handles all scaling and provisioning behind the scenes.
Serverless analytics with DuckDB and MotherDuck
DuckDB is an in-process analytical engine with no server to run at all — it executes queries directly inside your application process, which makes it a natural fit for serverless functions and ephemeral compute. MotherDuck extends DuckDB into a fully serverless cloud data warehouse: storage and compute are separated, compute scales automatically and idles to zero when you are not querying, and you are billed for what you use rather than for an always-on cluster. The result is a serverless analytics experience where you can query gigabytes to terabytes without sizing, patching, or babysitting infrastructure.
Related terms
data pipeline →
A data pipeline is a series of interconnected processes that extract data from various sources, transform it into a usable format, and load it into a…
query engine →A query engine parses, optimizes, and executes SQL queries against data sources. Learn its core architecture, how it differs from a database, and see examples like DuckDB and Presto.
in-memory database →An in-memory database is a type of database management system that primarily relies on a computer's main memory (RAM) for data storage and processing, as…
Cloud Storage →Cloud storage refers to a model of data storage where digital information is kept on remote servers accessed through the internet, rather than on local hard…
Online Analytical Processing (OLAP) →A complete guide to Online Analytical Processing (OLAP). Learn about OLAP cubes, the differences between OLAP and OLTP, and the types of OLAP systems (MOLAP, ROLAP, HOLAP).
ETL →ETL (Extract, Transform, Load) is a data integration process that combines data from multiple sources into a single destination, typically a data warehouse…
FAQS
DuckDB itself is in-process — it runs inside your application with no server or cluster to manage, which makes it an excellent fit for serverless functions. MotherDuck builds on DuckDB to provide a fully serverless cloud data warehouse, separating storage from compute and scaling automatically.
A serverless data warehouse separates storage from compute and provisions query resources on demand, scaling automatically (often to zero when idle) and billing based on usage instead of a fixed, always-on cluster. You load and query data without sizing or managing infrastructure.
A traditional database runs on servers you provision and keep running — and pay for — around the clock. A serverless database abstracts that away: compute scales up and down automatically with demand, can idle to zero, and is billed per use, so you never manage or pay for idle capacity.
The main trade-offs are cold starts (a short delay when compute spins up from idle) and reduced control over the runtime environment. For bursty or intermittent workloads these are usually outweighed by not paying for idle capacity and not having to manage scaling.
